centrail

The system of record for AI work.

Agent economics for engineering teams. Track AI coding spend across Claude Code, Codex, and GitHub Copilot CLI, with usage and estimated cost connected to your Git history.

Works with
CLI logs
  • Claude Code
  • Codex
  • GitHub Copilot CLI
Chat exports
  • ChatGPT
  • Claude
  • Gemini
Git evidence
  • GitHub
Ground level, not cloud level

Usage totals are only the beginning.

Your model provider shows consumption. Centrail brings supported coding-agent usage and Git context into the same record.

The consumption question

“How much AI did we use?”

Tokens, model totals, and spending without the engineering context.

The work question

“Where did that usage go?”

Inspect captured usage alongside repositories, matched commits, and the work that remains unattributed. Teams see only eligible shared usage.

See how to start your own record →
Why Centrail

Centrail sees what you don't.

Agent economics measures what AI agents consume, what they cost, and the work associated with that usage. Centrail starts with coding agents and Git.

Usage connected to work

See tokens and estimated cost alongside repositories and commits. Git attribution allocates captured usage; it does not prove an agent caused a commit.

Spend with context

Inspect AI coding cost by developer, repository, and model. Keep unattributed usage visible rather than forcing every token into a commit.

Teams with boundaries

Members choose which repositories to share with a team and whether to include earlier history. Installing the GitHub App does not share anyone's usage.

Questions engineering teams ask

How can I track Claude Code usage across a team?

Start with a free account and sync local agent logs. When you create a paid team, invite members to connect their devices and explicitly share the repositories the team should see. How team numbers work →

Can I track Copilot and Codex spend without a proxy?

Centrail reads supported local Claude Code, Codex, and GitHub Copilot CLI logs. No proxy, gateway, or application SDK is required. You still connect the Centrail CLI to an account; coverage depends on the logs available on each device. How local capture works →

Does cost per commit tell me what a feature cost to build?

It is an attributed estimate, not a feature invoice or proof of business value. Use it alongside repository context, model mix, and usage that has not been attributed to a commit. How attribution works →

How does Centrail estimate carbon, water, and electricity use?

GreenAI estimates energy from captured tokens and model assumptions, then derives water use and CO₂ with grid factors. These are footprint estimates, not your utility bill. Read the full methodology →

Read all frequently asked questions →

Get started · Free

Every token has a footprint.

Start your record of AI work. Connect your coding agents to see captured usage, estimated spend, and environmental footprint alongside the Git work associated with it.

01

Create a free account

No card required. Start with your own device before bringing a team.

02

Connect and sync

Follow the CLI guide and choose which repositories may sync.

03

Explore your first numbers

Check captured usage, estimated cost, Git attribution, and the GreenAI footprint.

GreenAI models energy, water, and CO₂ from captured usage. These are estimates, not measurements from a provider's data center. The methodology is public.

Proof, not promises

Real numbers. Honest pricing.

Start with your own usage for free. Bring your team when you're ready. Pricing below comes from the same plan policy as our pricing page.

hover or tap a number for what it equals

Tokens are self-reported: each account's Centrail CLI reads them from its own agent logs, and chat-app imports are estimated from exports. Energy, water and CO₂ are estimates, shown with the range their sources span. How they're estimated

Every account makes the picture sharper. Join the pledge and add your numbers, free.

“Arguably the most important change will come from speeding up the transition to clean energy, and boosting the amount available so that demand for greener AI does not gobble up the low-carbon electricity also needed to decarbonise other sectors, from transportation to construction.”
The Economist, “The tricky task of calculating AI’s energy use” (2025)
Free
$0
1 device, full history
Pro
$6/mo
3 devices, full history
Team
$49/mo
5 members included

Personal use is free, with no card required to begin. Sync first, inspect your numbers, then decide whether a paid team fits your workflow.

Compare team plans · Get started guide
Want a demo with Brayden? Contact support@centrail.org.